ASPIRE

Status: Plan

Analog Solvers for Physics-based Inverse Real-time Estimation. ASPIRE builds Analog PDE Processing Units (APPUs) to solve complex field equations with 100x the energy efficiency and significant speedup over modern GPUs.

Analog Computing Inverse Problems PDE Solvers Energy-Efficient HPC CMOS VLSI Scientific Hardware Real-time Inversion Edge Computing

ASPIRE (Analog Solvers for Physics-based Inverse Real-time Estimation) is a research project focused on building Analog PDE Processing Units (APPUs) to overcome the energy bottlenecks of modern digital scientific computing. Unlike traditional CPUs or GPUs that discretize physics into binary logic, ASPIRE maps mathematical field equations directly onto physical silicon circuits.

This approach, termed “Physical Emulation,” enables real-time, low-power solving of field equations, with long-term targets including complex fluid dynamics (Navier-Stokes) and potential applications in Poisson solvers.

Primary Innovation

Solving Partial Differential Equations (PDEs) through physical emulation on gridded problems like finite differences rather than numerical simulation.

Impact & Applications

  • Geosciences
  • Biomedical

Key Technologies

  • Analog PDE Processing Units (APPUs)
  • VLSI / CMOS Design
  • Translinear Circuits
  • Hybrid Analog-Digital Solvers

Target Impact

  • 100x energy improvement and significant speedup over modern GPUs.

Team Members

Collaborators